Low-cost and secure multi-image encryption scheme based on P-tensor product compressive sensing

Abstract

In order to ensure security, the existing image encryption schemes based on traditional compressive sensing (CS) need to generate different measurement matrices according to different images. Due to the requirement of “dimension matching” in matrix multiplication, the larger these sampled images are, the larger these measurement matrices are. These will seriously consume the storages of the communicating parties. In view of the above problems, this paper proposes a double-layer multi-image encryption scheme with lower storage consumption and higher security based on P-tensor product CS. The scheme makes full play to the advantages of P-tensor product in matrix multiplication, and uses the same basic measurement matrix to process multiple images of different sizes. Some intermediate cipher images of different sizes are then generated. In order to ensure security, the basic measurement matrix is updated to generate the needed measurement matrices for other images. Finally, the intermediate cipher images of different sizes are fused and scrambled to one final cipher image. The experiments and analyses show that our scheme greatly reduces the storage consumption while ensuring security.